Example 12 illustrates the ease of model fitting when

Info icon This preview shows pages 9–11. Sign up to view the full content.

Example 1.2 illustrates the ease of model fitting when binary-valued covariates are known; the notion of permutation similarity plays a similar role in the case of k -ary covariates. 1.3.2. Approximate inference The careful reader will have noted that in the case of known categorical covariates, examples such as those above can be expressed as contingency tables – a notion we revisit in Section 1.4 – and hence may admit exact inference procedures. However, if covariates are latent, then an appeal to maximum-likelihood estimation induces a combinatorial optimization prob- lem; in general, no fast algorithm is known for likelihood maximization over the set of covariates and Bernoulli parameters under the general k -group stochastic block model. The principal difficulty arises in maximizing the n -dimensional k -ary covariate vector c over an exponentially large model space; estimating the Copyright © 2014. Imperial College Press. All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law. EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 2/16/2016 3:37 AM via CGC-GROUP OF COLLEGES (GHARUAN) AN: 779681 ; Heard, Nicholas, Adams, Niall M..; Data Analysis for Network Cyber-security Account: ns224671
Image of page 9

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

10 B. P. Olding and P. J. Wolfe Fig. 1.3. Representations A and Π A Π of data drawn from the stochastic block model of Example 1.3, corresponding to isomorphic graphs (black boxes denote links). Though p 01 = 0, only a small subset of permutation similarity transformations Π ( · ) Π will reveal the disconnected nature of this network. ( k +1 2 ) associated Bernoulli parameters then proceeds in exact analogy to Example 1.2 above. The following example illustrates the complexity of this inference task. Example 1.3 (Permutation and Maximization). Consider a 100- node network generated according to the stochastic block model of Defini- tion 1.2 , with each group of size 50 and p 00 = p 11 = 1 / 2 , p 01 = 0 . Figure 1.3 shows two permutation-similar adjacency matrices, A and Π A Π , that cor- respond to isomorphic graphs representing this network ; inferring the vector c of binary categorical covariates from data A in Figure 1.3a is equivalent to finding a permutation similarity transformation Π A Π that reveals the distinct division apparent in Figure 1.3b. Given the combinatorial nature of this problem in general, it is clear that fitting models to real-world network data can quickly necessitate approximate inference. To this end, Example 1.3 motivates an important means of exploiting algebraic properties of network adjacency structure: the notion of a graph spectrum . Eigenvalues associated with graphs reveal several of their key properties (Chung, 1997) at a computational cost that scales as the cube of the number of nodes, offering an appealing alternative in cases where exact solutions are of exponential complexity.
Image of page 10
Image of page 11
This is the end of the preview. Sign up to access the rest of the document.
  • Spring '12
  • Kushal Kanwar
  • Graph Theory, Statistical hypothesis testing, Imperial College Press, applicable copyright law

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern